【作者】 曹佃省；

【导师】 谢光荣；

【作者基本信息】 中南大学 ， 精神病学与精神卫生， 2010， 博士

【摘要】 第一部分论文理论背景、研究现状及选题意义该部分概括了健康行为与健康、影响青少年健康锻炼行为的因素及国内青少年健康锻炼行为的研究现状,概括了健康心理学领域的连续模型及阶段模型,阐明了研究选题的现实意义及理论基础。第二部分健康行为程式模型(HAPA模型)适用性检验研究目的检验HAPA模型在中国青少年健康锻炼行为中的适用性,探求不同社会心理变量在不同锻炼阶段间的预测作用以利于后续分组实验干预研究：检验自我效能、危险认知、结果预期是否能有效的预测行为意向；检验行动计划在行为意向向锻炼行为转化过程中的中介效应是否存在；检验HAPA行为阶段在青少年健康锻炼行为中的存在性及模型中的社会心理变量在不同行为阶段问的阶段差异性。对象与方法本部分研究采用分层整群随机取样法选取706名初高中青少年为研究对象,对HAPA模型中的构成要素如危险认知、结果预期、自我效能、行动计划及锻炼活动等因素采用相应量表收集数据。结果与讨论(1)关于自我效能、危险认知、结果预期等变量对行为意向的预测作用自我效能(β=.43,p<.001)、危险认知(β=-.10,p<.001)及锻炼消极预期(β=-.08,p<.05)显著预测行为意向,其中锻炼自我效能的预测作用最为显著；积极预期觉知(β=.01,p>.05)不是行为意向的显著预测变量；(2)关于行动计划在行为意向与锻炼行为间的中介效应中介效应分析证明行动计划是行为意向与锻炼行为间的中介变量(Sobel Z=8.82,p<.001),表明行为意向到锻炼行为的转化过程中,部分经由行动计划起作用,与HAPA模型的理论假设一致；(3)关于行为阶段存在性的检验多元逐步回归分析结果显示,自我效能、危险认知、锻炼积极预期觉知、消极预期觉知、行动计划等变量在不同行为阶段间对锻炼活动的解释程度不同,结果支持HAPA中所假定的行为阶段的存在。在前意向阶段,自我效能(β=.48,p<.001)、危险认知(β=-.21,p<.05)、积极预期觉知(β=.21,p<.05)是锻炼行为的显著预测变量,解释40%的行为变异量(R2=0.40)；锻炼消极预期(β=-.01,p>.05)及行动计划(β=.06,p>.05)不是显著预测变量。在意向阶段,自我效能(β=.37,p<.001)及行动计划(β=.16,p<.01)是锻炼行为的显著预测变量,解释27%的行为变异量(R2=0.27)；危险认知(β=-.05,p>.05)、积极预期觉知(β=.02,p>.05)、消极预期觉知(β=-.07,p>.05)不是锻炼行为的显著预测变量。在行动阶段,危险认知(β==-.20,p<.001)、自我效能(p=.17,p<.05)、行动计划(β=.28,p<.001)是该阶段锻炼行为的显著预测变量,解释20%的行为变异量(R2=0.20)；积极预期觉知(p=-.01,p>.05)及消极预期觉知(p=.01,p>.05)不是显著预测变量。(4)关于社会心理变量在不同行为阶段间差异显著性的检验：自我效能(F(2,696)=44.54,p<.01)、危险认知(F(2,681)=17.30,p<.01)、消极预期觉知(F(2,684)=20.64,p<.01)、行动计划(F(2,680)=97.82,p<.01)及锻炼活动(F(2,694)=86.52,p<.01)等变量在前意向者、意向者及行动者三阶段间存在显著差异；行动者自我效能最高,锻炼积极预期觉知最多,参与的锻炼活动最多。前意向者及意向者间在危险认知水平(t(430)=.52,p=.60)及消极预期感知(t(432)=2.42,p=.06)等方面差异不显著。第三部分基于HAPA模型的健康行为分组纵贯实验干预研究研究目的检验行为阶段匹配或非匹配干预在不同行为阶段群体中的效用性,检验是否阶段匹配干预效应最大而非匹配干预效应弱化。研究假设危险认知交流干预对前意向阶段者是匹配干预、计划策略制定干预对意向阶段者是匹配干预,匹配干预能显著促进相应阶段内个体的意向水平及锻炼行为；假定危险认知或计划策略干预对行动阶段者皆为非匹配干预,不能引起行为意向或锻炼行为的显著改变。对象与方法分层整群选取初高中青少年被试,初测693名被试分层整群随机化分为危险认知交流组、计划策略制定组及对照组,四周后534名被试完成复测。采用3×3×2组间因素纵贯干预设计,初测复测行为意向及锻炼行为等数据,采用重复测量设计方差分析方法处理数据。结果与讨论(1)锻炼意向干预前后不同阶段不同组别间的分析对整体样本：组间分析结果显示存在时间主效应(F(2,525)=35.07,p<.001,η12=.12),干预前后锻炼意向改变显著；不存在组间分组处理主效应(F(2,525)=.83,p=.44)或处理x锻炼阶段交互效应(4,525)=1.03,p=.39)。组内分析结果显示存在组内时间主效应(F(1,525)=10.41,p<.001,η2=.02),时间x处理交互效应(F(2,525)=7.02,p<.10,η2=.03),时间x锻炼阶段交互效应(F(2,525)=5.41,p<.005,η2=.02),显示锻炼意向水平干预后比干预前增加,不同干预分组、不同锻炼阶段间获益改变程度有高低不同,与研究假设一致。对危险认知交流组：组间分析结果显示行为意向存在锻炼阶段主效应(F(2,86)=6.51,p<.01,η2=.13),显示意向在不同锻炼阶段间存在差异,前意向者锻炼意向水平最低,行动者意向水平最高。组内分析结果显示存在时间主效应(F(1,86)=19.25,p<.001,η2=.18),显示锻炼意向水平干预后比干预前增加。前意向者、意向者及行动者在危险认知干预后锻炼意向水平都有所增加。前意向者行为意向增幅程度最大,与匹配干预获益最大的假设一致；意向者及行动者干预后行为意向也显著增强,与研究假设不一致,结果表明危险认知交流干预在改变个体锻炼意向方面在各行为阶段间具有普遍效用。对计划策略干预组：组间分析结果显示存在行为阶段主效应(F(2,220)=17.45,p<.001,η2=.14),显示行动阶段个体的行为意向水平最高,前意向阶段的意向水平最低；组内分析结果显示行为意向存在时间x行为阶段交互效应(F(2,220)=3.52,P=.03,η2=.03),表明行动计划干预在不同的锻炼阶段间对行为意向的影响程度不同。结果显示前意向阶段及意向阶段都从计划策略干预获益。干预后意向阶段者的行为意向增加显著,行动阶段者意向水平没有变化,符合匹配干预获益最大的假定；前意向者的行为意向在计划干预后也显著增加,与研究假设不一致。对照组：只存在组间锻炼阶段主效应(F(2,219)=21.57,p<.001,η2=.17),没有时间主效应或时间x阶段或时间x处理交互效应,显示对照组在研究前后时间内,行动阶段者依然较前意向者或意向者参与锻炼的意向水平高,各阶段间锻炼意向都没有随时间发生改变。(2)锻炼频率干预前后不同阶段不同组别间的分析对整体样本：组间分析结果显示锻炼行为存在组间锻炼阶段主效应(F(2,525)=36.50,p<.001,η2=.12),行动阶段者锻炼活动最多,不存在处理主效应(F(2,525)=.14,p=.87)或行为阶段x处理交互效应(F(4,525)=3.39,p=.85)；组内分析结果显示存在时间主效应(F(1,525)=13.76,p<.001,η12=.03)、时间x处理交互效应(F(2,525)=2.36,p<.10,η2=.01)、时间x阶段交互效应(F(2,525)=11.24,p<.001,η2=.04)、时间x阶段x处理交互效应(F(4,525)=2.52,p<.05,η2=.02),表明锻炼行为干预前后在不同阶段、不同干预分组间的获益改变程度存在差异。对危险认知交流组：组间分析结果显示存在组间锻炼阶段主效应(F(2,86)=6.05,p<.01,η2.12),表明行动阶段者锻炼活动最多,前意向阶段者锻炼活动最少；组内分析结果显示存在组内时间主效应(F(1,86)=7.14,p<.01,η2.08)、时间x锻炼阶段交互效应(F(2,86)5.21,p<.01,η2.11),显示干预后锻炼行为显著增加,增加程度在不同阶段间存在差异,前意向阶段者获益增加最多,与匹配干预获益最大的假设一致。对计划策略干预组：组间分析结果显示存在锻炼阶段主效应(F(2,220)=17.61,p<.001,η2=.14),行动者锻炼活动最多,前意向者锻炼活动最少。组内分析结果显示存在组内时间主效应(F(1,220)=9.08,p<.01,η2=.04)、时间x锻炼阶段交互效应(F(2,220)=10.14,p<.001,η2=.08),显示干预后锻炼行为显著增加,增加程度在不同阶段间存在差异,前意向阶段与意向阶段者获益增加最多。意向阶段者计划策略干预后锻炼行为显著增加与匹配干预获益最大的假设一致,前意向阶段者干预后锻炼行为也显著增加与研究假设不一致。对照组：组间锻炼阶段主效应(F(2,219)=30.80,p<.001,η2.22),不存在时间主效应(F(1,219)=.98,p>.10,η2<.01)或时间x锻炼阶段交互效应(F(2,219)=2.50,p>.05,η2=.02),表明对照组在研究前后时间内,行动阶段者依然较前意向或意向阶段者锻炼活动要多,各阶段间的锻炼行为没有随时间发生改变。第四部分健康锻炼行为养成中的中介-调节模型构建研究目的检验锻炼动机及锻炼技能在行为意向向行为活动转化中调节效应是否存在,探究行为意向向行为活动转化过程中的中介调节机制。对象与方法分层随机选取初一到高三的青少年被试,初测693名,4周后复测534名。采用相关量表测量行为意向、锻炼动机、行动计划、运动技能及锻炼行为等变量。分析中缺失数据用最大似然法加以估计,利用Preacher主张的调节效应检验方法,检验内在动机、运动技能等在行动意向、行动计划与锻炼行为间的调节效应是否存在。结果与讨论在行为意向、锻炼动机、行动计划、运动技能与锻炼行为间：(1)行动计划是行为意向与行为活动间的中介变量；(2)内在动机是行为意向与行动计划间的调节变量,行为意向向行动计划的转化受内在动机水平的调节,内在动机水平高,越能够制定更多的行动计划,引起更多的锻炼行为；(3)运动技能是行动计划与锻炼行为间的调节变量,行动计划转化为锻炼行为受个体运动技能多少的调节,运动技能越多者,越能将行动计划转变成锻炼行为。第五部分研究概括结论(1)横向研究结果证明HAPA模型主张的行为阶段在青少年锻炼行为中存在,HAPA模型构成要素在不同行为阶段间存在显著差异,HAPA模型适用于青少年健康锻炼行为。(2)纵贯研究结果部分支持依据HAPA阶段模型理论所进行的实验干预,证明了匹配干预的优越性。(3)锻炼技能、动机类型是行为意向到锻炼行为转化间的重要调节变量,内在动机水平高、锻炼技能多者促进行动计划的主动制定及有效执行。更多还原

【Abstract】 Part 1 General Introduction of Research BackgroundIn this part, an overview on associations of health behaviors with psychological and physical health and factors affecting health behaviors during adolescence were addressed. The current study background in adolescent health behaviors was presented and a summarization about health behavior models and theories were demonstrated in this part.Part 2 Feasibility and Applicability Examination on HAPAObjective To test HAPA model’s validity and feasibility among Chinese adolescents in the domain of physical activity; To test variables predicting intention; To test planning’s effect between intention and behavior; To test whether social cognitive variables’ stage difference exists or not.Participants and MethodsStratified sampling method was conducted.706 participants from middle and high school provided valid data. Constructs in HAPA (risk perception, perceived pros and cons, self-efficacy, action planning) were measured. One-way ANOVA and mediation analysis were conducted to test whether social cognitive variables’ stage difference and plannings’ mediating effect exists or not.Results and Discussion(1) Results of Variables Predicting IntentionSelf-efficacy (β=.43, p<.001), risk perception (β=-.10, p<.01) and perceived cons (β=-.08, p<.05) were significant predictor on intention and self-efficacy was the strongest predictor among all the variables. Perceived pros (β=.01, p>.05) was not a significant predictor on intention for adolescent PA behavior, indicating that even adolescents perceived the benefits (pros) of regular activity, they yielded to the perceived barriers (cons) and with weak intention for PA participation.(2) Results of Planning Mediating Analysis for Intention, and BehaviorMediating analysis demonstrated that planning was a partial mediator between intention and behavior, intention’s indirect effect on behavior was partially mediated by planning, consistent with the assumption of HAPA.(3) Results of Stage CheckingHierarchical Regression analysis results revealed that self-efficacy, risk perception, perceived pros and cons and planning showed different predicting effect on physical activity, supporting the stage presumption in HAPA. Self-efficacy was a universal strong predictor for all the three stages. In preintention stage, self-efficacy (β=.48, p<.001), risk perception (β=-.21, p<.05), perceived pros (β=.21, p<.05) were significant predictor of behavior except perceived cons (β=-.01, p>.05) and planning (β=.06, p>.05), explaining 40%variance of PA change (R2= 0.40).In intention stage, self-efficacy (β=.37, p<.001) and planning (β=.16, p<.01) were significant predictor of behavior except risk perception (β=-.05, p>.05), perceived pros (β=.02, p>.05) and cons (β=-.07, p>.05), explaining 27%variance of PA change (R2= 0.27).In action stage, self-efficacy (β=.17, p<.05), risk perception (β==-.20, p<.001) and planning (β=.28, p<.001) were significant predictor on behavior except perceived pros (β=-.01, p>.05) and cons (β=.01, p>.05), explaining 20%variance of physical activity change (R2= 0.20).(4) Results of Social Cognitive Variables’Stage DifferenceSelf-efficacy (F(2,696)= 44.54, p<.01), risk perception (F(2,681)= 17.30, p<.01), perceived cons (F(2,684)= 20.64, p<.01), action planning (F(2,680)= 97.82, p<.01) and physical activity (F(2,694)= 86.52, p<.01) all showed significant stage difference among preintenders, intenders and actors. Actors had the highest level of self-efficacy, perceived the highest level of pros and the lowest levels of risk perception, had more planning and participated in the most amount of physical activity, no stage difference was found between actors and intenders in perceived pros. Compared to actors, preintenders and intenders perceived higher level of cons. No stage difference was found between preintenders and intenders as to risk perception (t(430)=.52, p=.60) and perceived cons (t(432)= 2.42, p=.06).Part 3 Longitudinal Intervention Study Based on HAPAObjective To test the effectiveness of stage-matched or mis-matched intervention on physical activity change. We hypothesize that risk resource communication intervention would be matched to preintenders and preintenders would benefit most from resource communication intention; Action planning intervention would be matched to intenders and intenders would benefit most from planning intervention; Actors would not benefit from neither risk resource communication nor action planning making intervention for neither of them was matched to actors.Participants and MethodsStratified sampling method was used.693 adolescents completed baseline questionnaire and were randomly designed into resource communication group, strategic planning group and control group. Students in the treatment groups received a one-hour lesson and were given printed materials.534 adolescents completed the follow-up study. PA stage, PA intention and PA frequency were measured at the two time point. A quasi-experimental 3x3x2 between-factors design with repeated measures was chosen.3 stages of change (preintenders, intenders, actors) and three intervention groups (resource communication, planning, control). PA frequency and PA intention were measured at baseline and follow-up and served as the dependent variable.Repeated Measures Analyses of Variance was conducted with stages (preintender, intender, actor) and treatment (risk resource communication, planning, control group) were chosen as between-subjects factors and PA frequency and PA intention were run separately as the dependent variable measured at two points in time four weeks apart (pre-post measures). Missing data were imputed using the Expectation Maximization (EM) algorithm in SPSS.Results and Discussion(1) Repeated Measures ANOVA on PA intention before and after Intervention among Different Groups and StagesFor all the participants:Between group analysis results showed the existence of time effect (F(2,525)= 35.07, p<.001,η2=.12), indicating the significant change of PA intention before and after intervention, no treatment effect (F(2,525)=.83, p=.44) or treatment x stage interaction effect (F(4,525)= 1.03, p=.39) was found, for the whole sample participants consisting of different stages with preintenders, intenders and actors, no intervention was targeted to match the whole group, the results was consistent with the hypothesis; Within group analysis results showed the existence of time effect (F(l,525)=10.41, p<.001,η2=.02), time x intervention interaction effect (F(2,525)= 7.02, p<.10,η2=.03), time x stage interaction effect (F(2,525)= 5.41, p<.005,η2=.02), indicating the existence of significant change of PA intention before and after intervention differed among different intervention groups and stages.Risk Resource Communication Group:Between group analysis results showed a stage effect (F(2,86)= 6.51, p<.01,η2=.13), revealing the significant difference of PA intention between different stages. Within group analysis results showed a time effect (F(1,86)= 19.25, p<.001,η2 =.18), preintenders, intenders and actors all increased their PA participation intention whereas preintenders and intenders benefited most and had the steepest increase. The increase of PA intention for preintenders was consistent with the hypothesis that stage-matched interventions had superiority over mis-matched ones whereas the increase for intenders and actors was unexpected.Planning Group:Between group analysis results showed a stage effect (F(2,220)= 17.45, p<.001,η2=.14), revealing that preintenders had the lowest levels of PA intention whereas actors had the highest level. Within group analysis showed a time x stage interaction effect (F(2,220) = 3.52, p=.03,η2=.03), revealing that planning intervention had different influence on different stages. Preintenders and intenders both benefited from planning intervention, the increase in intenders was consistent with the hypothesis whereas the increase in preintenders was unexpected.Control Group:Only a between group stage main effect (F(2,219) = 21.57, p<.001,η2=.17) was found, no time main effect or time x stage or time x intervention interaction effect was found, revealing that actors remained the highest levels of PA intention than preintenders and intenders, there was no PA intention change in all the 3 stages in control group during the two time point study.(2) Repeated Measures ANOVA on PA frequency before and after Intervention among Different Groups and StagesFor all the participants:Between group stage effect (F(2,525)= 36.50, p<.001,η2=.12), no treatment effect (F(2,525)=.14, p=.87) or stage x treatment interaction effect (F(4,525)= 3.39, p=.85), for the whole sample participants consisting of different stages with preintenders, intenders and actors, no intervention was targeted to match the whole group, the results were consistent with the hypothesis; Within group analysis results showed the existence of time effect (F(1,525)= 13.76, p<.001,η2=.03), time x treatment interaction effect (F(2,525)= 2.36, p<.10,η2=.01), time x stage interaction effect (F(2,525)= 11.24, p<.001,η2=.04), time x stage x treatment interaction effect (F(4,525)= 2.52, p<.05,η2=.02), demonstrating the existence of significant change of PA participation frequency before and after intervention, the change differed across different intervention groups and stages.Resource Communication Group:Between group analysis results showed the existence of stage effect (F(2,86)= 6.05, p<.01,η2=.12); Within group analysis showed a time effect (F(1,86)= 7.14, p<.01,η2 =.08), time x stage interaction effect (F(2,86)= 5.21, p<.01,η2=.11), revealing that the change of PA frequency differed among different stages and different groups before and after the intervention. Preintenders had the steepest increase, consistent with the hypothesis that stage-matched interventions had superiority over mis-matched ones.Planning Group:Between group analysis results showed the existence of stage effect (F(2,220)= 17.61, p<.001,η2=.14). Within group analysis results showed the existence of time effect (F(l,220)= 9.08, p<.01,η2=.04), time x stage interaction effect (F(2,220)= 10.14, p<.001,η2=.08), revealing the significant increase of PA frequency among different stages. Preintenders and intenders showed the steepest PA participation increment. The steep increment in intenders was consistent with the hypothesis that stage-matched interventions had superiority over mis-matched ones whereas the increase in preintenders was unexpected.Control Group:Between group analysis showed the existence of stage effect (F(2,219)= 30.80, p<.001,η2=.22), no time effect (F(1,219) .98, p>.10,η2<.01) or time x stage interaction effect (F(2,219)= 2.50, p>.05,η2=.02), revealing that actors remained the highest levels of PA participation than preintenders and intenders, there was no PA behavior change in all the 3 stages in control group during the two time point.Preintenders showed unexpected increment both in PA intention and behavior change after strategic planning intervention, inconsistent with the study hypothesis. The unexpected intervention effect might be that before intervention, preintenders perceived highest levels of cons and the lowest levels of exercise efficacy and the least amount of planning and stayed in the no intention no action stage. After planning intervention, preintenders increased their planning making strategies and exercise self-efficacy, resulting in the enhancement in PA participation intention and the change in behavior.Part 4 Moderated-Mediation Model in Health Behavior FormationObjective To examine whether the moderated mediation mechanism exist or not in the translating process from intention to behavior change: whether the type of exercise motivation (intrinsically-motivated or extrinsically-motivated) and exercise skill’s moderating effect exist or not among intention, planning and behavior change process.Participants and Methods 693 adolescents from middle and high shcool completed the baseline study and 534 of them completed the follow up study. PA intention, exercise motivation (BREQ), exercise skill and physical activity (IPAQ) were measured at baseline. Planning and PA were measured at follow-up four weeks later.Results and DiscussionIn the relation between intention, exercise motivation, exercise skill, action planning and behavior:(1) Mediation Model:Action planning was a partial mediator between PA intention and behavior, the translation from intention to behavior partially came from the effect of action planning;(2) Moderating Mediation Model:Exercise motivation as a ModeratorIn the mediating model, exercise motivation was a moderator between intention, action planning, and behavior, one who had higher levels of intrinsic motivation was apt to form more action planning and thus tended to have more levels of behavior change.(3) Moderated Mediation Model:Exercise Skills as a ModeratorIn the mediating model, exercise skill was a moderator for the translation from action planning making to behavior change, one who grasped more types of exercise skills was apt to execute their action planning and thus tended to have more levels of behavior change. Part 5 General Conclusion(1) Cross-sectional study revealed the existence of stages and social cognitive variables’stage difference in HAPA. The feasibility of HAPA model was validated in the study of physical activity behavior among adolescents.(2) Longitudinal study results partly support the hypothesis that stage-matched intervention had superiority over mis-matched ones: intervention matched to a specific stage had more efficacy than mis-matched to any other stages.(3) Exercise skill and exercise motivation are important moderator for the transition between intention, action planning and behavior change.更多还原